Validation of object oriented metrics using open source software system

Abstract

In today's environment the relevance of Free Open Source Software Systems is understood and appreciated both in academia and research. The debate about the pros and cons of the open source vis-à-vis proprietary software has been raging from ages ever since Richard Stallman founded the Free Software Foundation in 1985. With the changing trends in the domain of Object Oriented Systems there is a need to measure the fault predictability of software metrics on open source software systems. In this paper we present the results of empirical study which was conducted using open source software, JHotDraw 7.5.1. We computed the object oriented metrics, proposed by Chidamber and Kemmerer, and performed bug-class mapping for the software under study. We also studied the relationship between the revisions made to open source software and its software metrics measure.

State-of-the-Art in Empirical Validation of Software Metrics for Fault Proneness Prediction: Systematic Review

“…Interestingly, none of the studies found RFC insignificant but 5 of the studies did not measure RFC. [10], [19], [21], [17], [1], [22], [6], [25], [23], [3], [4], [2], [24], [27], [11], [29], [12], [9], [30], [31], [32], [36], [37] [8] [5], [28], [35], [34], [33] RFC [10], [19], [21], [17], [1], [22], [6], [11], [23], [3], [4], [8], [5], [24], [27], [11], [29], [12], [9], [30], [31], [32], [36], ...…”

“…With the insignificance results, it show that a class having a higher number of children is not likely to be fault-prone than others with less number of children. The validation for both DIT and NOC are shown in Table 6 [19], [22], [6], [25], [5], [2], [27], [9], [37] [10], [21], [17], [23], [3], [4], [8], [24], [11], [29], [12], [30], [31], [32], [36] [1], [28], [35], [34], [33] NOC [17], [22], [2] [19], [25], [23], [3], [4], [24], [27], [11], [29], [12], [30], [31], [32], [36], [37] [10], [21], [1], [6], [5],…”

“…(See Table 7) The implication of the results is that a class having a larger number of lines of code is more likely to have faults than classes with small code lines. [17], [1], [22], [25], [3], [4], [2], [24], [27], [28], [11], [29], [30], [31], [32], [33], [36] - [10], [19], [21], [6], [23], [8], [5], [12], [35], [34], [37] **N/A= not applicable…”

Untitled

“…However, 22 studies confirmed WMC significance relationship with the FP of OO classes, one study [2] found considered it to be insignificant while 6 others studies did not consider it their studies.This is captured in Table 3. [10], [40], [17], [1], [22], [6], [25], [23], [3], [4], [8], [2], [24], [27], [11], [29], [12], [31], [32], [35], [36], [37] [2]…”

A Systematic Review of the Empirical Validation of Object-Oriented Metrics Towards Fault-Proneness Prediction

“…In the same vein, 15 studies found it to be insigni¯cant while 5 studies did not measure [11], [28], [4], [5], [29], [30], [34], [7] [ 12], [8], [9], [27], [45] [31], [32], [38], [6], [7], [33], [39], [13], [41], [14], [43], [44], [47], [48] it. With more emphasis, the insigni¯cance of DIT indicates that a class with higher number of inheritance hierarchy is not likely to have faults.…”